Spacecraft Radiator Freeze Protection Using a Regenerative

Heat Exchanger with Bypass Setpoint Temperature Control

Eugene K. Ungar
NASA/Johnson Space Center
Houston, TX 77062

submitted for
38th International Conference on Environmental Systems
June 29-July 3, 2008
Hyatt Regency
San Francisco, California, USA
‘

Abstract
Spacecraft radiators are sized for their maximum heat load in their warmest
thermal environment, but must operate at reduced heat loads and in colder
environments. For systems where the radiator environment can be colder than
the working fluid freezing temperature, radiator freezing becomes an issue.
Radiator freezing has not been a major issue for the Space Shuttle and the
International Space Station (ISS) active thermal control systems (ATCSs)
because they operate in environments that are warm relative to the freezing point
of their external coolants (Freon-21 and ammonia, respectively).
For a vehicle that lands at the Lunar South Pole, the design thermal environment
is 215K, but the radiator working fluid must also be kept from freezing during the
0 K sink of transit. A radiator bypass flow control design such as those used on
the Space Shuttle and ISS requires more than 30% of the design heat load to
avoid radiator freezing during transit - even with a very low freezing point working
fluid. By changing the traditional ATCS architecture to include a regenerating
heat exchanger inboard of the radiator and by using a regenerator bypass flow
control valve to maintain system setpoint, the required minimum heat load can be
reduced by more than half. This gives the spacecraft much more flexibility in
design and operation.
The present work describes the regenerator bypass ATCS setpoint control
methodology. It includes analytical results comparing the performance of this
system to the traditional radiator bypass system. Finally, a summary of the
advantages of the regenerator bypass system are presented.

Background
Active thermal control systems (ATCSs) on human spacecraft provide fluid at a
controlled temperature to cool equipment, cool the crew cabin air, and provide
condensation control. When spacecraft radiators are the major source of heat
rejection, as was the case on the Apollo Service Module, is the case on the
Space Shuttle Orbiter and the International Space Station (ISS), and will be the
case on the upcoming Crew Exploration Vehicle (CEV), their coolant return
temperature is normally set by controlling the radiator bypass flow. That is, a
preset coolant return temperature is maintained by mixing a variable amount of
warm coolant from the radiator inlet with the cold radiator outlet flow. All four
vehicles listed above use this approach.
Spacecraft radiators are sized for their maximum heat load in their warmest
thermal environment, but must operate at reduced heat loads and in colder
environments. For systems where the radiator environment can be colder than
the working fluid freezing temperature, radiator freezing becomes an issue.
Radiator freezing has not been a major issue for the Space Shuttle and ISS
ATCSs because they operate in environments that are warm relative to the
freezing point of their coolants (Freon-21 and ammonia, respectively). The
Apollo Service Module flew through extremely cold environments during its trip to
the Moon. The minimum ATCS heat load required to avoid freezing was reduced
to manageable values by use of a stagnation radiator - where the radiator fin
efficiency is greatly reduced at low heat loads by stagnating the flow through
much of the radiator.
For a vehicle that lands at the Lunar South Pole, the design thermal environment
is 215K1. The radiator working fluid must also be kept from freezing during the
0 K sink of transit. Because only safe working fluids can now be used in the crew
cabin, an Apollo-like single internal/external ethylene glycol/water loop is not an
option. If the allowed propylene glycol/water mixture is used in a stagnation
radiator, its higher freezing point results in an allowable heat load range greatly
reduced from Apollo's range of 65% to 100% of full load. At best, such a design
would have a very narrow operating range. At worst, such a design is not
feasible.
Even with a low freezing point working fluid in a traditional radiator bypass flow
control design, a large fraction of the design heat load would be required to avoid
radiator freezing during transit to the Moon. An alternate ATCS architecture was
developed to provide a flexible system for transit to and operation on the Moon.

System Design and Operation

Radiator Bypass Setpoint Control ATCS
Overview - A simplified radiator bypass thermal system is shown in Figure 1.
The standard two loop design includes an internal loop in the pressurized volume

1
Assuming that the ground slopes away from the vehicle at 5° and adding in the 1.54° tilt of the
lunar axis gives a worst case sun angle of 6.54°. The lunar surface optics are α/ε=0.865/0.95, so
a simple energy balance using a deep space (0 K) view and a solar constant of 1371 W/m2 yields
a Lunar surface temperature of 224 K. A vertical cylindrical radiator with a 10 mil silver-Teflon
coating (α/ε=0.10/0.85) has an average sink temperature of 215 K

2
plus an external loop. This way the internal loop can use a non-toxic fluid with
good heat transfer characteristics (like water) and the external loop can use a low
freezing point fluid that may be toxic. The two loops are connected by an
interloop heat exchanger. The loop flow rates are matched so that the product of
mass flow rate and specific heat is the same in the two loops, maximizing the
average radiator temperature and heat rejection capability. A radiator bypass
controls the mixed radiator fluid return temperature. A similar design is used in
the Space Shuttle Orbiter and in the international Space Station (ISS) ATCS.
The radiator performance in this type of system can be explored using a simple
analytical model.

Pressurized Volume
coldplate

coldplate

Accum.
H/X

heater
T pump

H/X

Accum.

coldplate coldplate
pump
T

radiators

Figure 1 –Active Thermal Control Design with Radiator Bypass

Radiator Analysis - The performance of the radiator bypass ATCS was explored
using a simplified Excel thermal model. The model predicts the radiator flow and
temperature distribution using a number of simplifying assumptions.
The differential local heat rejection, dQ, from a given differential area of radiator,
dA, is
dQ = εησdA Troot
4
(
− T∞4 ) (1)

3
where ε is the radiator emissivity, η is the radiator fin efficiency, σ is the Stefan-
Boltzman constant, Troot is the radiator root temperature, and T∞ is the radiator
environment temperature.
If the liquid flow tubes are evenly spaced in the radiator and have uniform flow,
the differential temperature drop for the radiator fluid, dT, is
dQ
dT = (2)
& cp
m
& is the radiator mass flow rate and cp is the radiator fluid specific heat.
where m
If
ƒ the fin root temperature is identical to the local fluid temperature
ƒ total radiator area is expressed as the product of its length, L, and width, w
ƒ ζ is defined as x/L, the non-dimensionalized length along the flow path
equations (1) and (2) can be combined as
(
& c p dT = εησwLdζ T 4 − T∞4
m ) (3)

or
dT εησA 4
dζ
=
& cp
m
T − T∞4( ) (4)

By representing the radiator mass flow rate as a fraction, Fradiator, of the full mass
& max , eqn. (4) becomes
flow rate, m
dT
=
εησA
dζ Fradiator m max c p
&
(
T 4 − T∞4 ) (5)

Making the further simplifications that

ƒ the radiator fin efficiency is constant
ƒ the fluid specific heat is constant
allows all the constants in eqn. (5) to be represented by a single constant, C
dT
=
C
dζ Fradiator
(
T 4 − T∞4 ) (6)

This equation allows the heat transfer to be tracked through the radiator by
following only the fluid temperature.
For the ATCS modeled here, the key temperatures are:
ƒ The setpoint temperature, Tsetpoint, is 283 K
o this is the mixed outlet temperature at all times and is the radiator
outlet temperature at full load in the design environment

4
 ƒ The external loop temperature drop at the design point, ΔTdesign, is 20 K
o this is the temperature difference between the radiator inlet
temperature and the setpoint temperature at all times
o notably, this is also the temperature rise across the internal loop at
full load
ƒ The design sink temperature is 215 K
The radiator performance is analyzed by calculating C for the design case (by
following the radiator fluid through the radiator). Once C is found, the radiator
can now be analyzed to calculate the radiator flow fraction, Fradiator, that yields the
desired radiator mix temperature at any heat load and in any environment. The
setpoint is maintained when the following equation is satisfied
Tsetpo int = (1 − Fradiator )Tradiator ,in + Fradiator Tradiator ,out (7)

where Tradiator,out is the calculated radiator outlet temperature. The radiator inlet
temperature, Tradiator,in is calculated from an energy balance
Tradiator ,in = Tsetpo int + ΔTdesignFload (8)

where Fload is the fraction of full load.

The radiator performance in the 215 K design sink is shown in Figure 2. The
figure shows that the radiator flow decreases non-linearly with heat load. The
flow fraction at 50% of peak load is less than 20%.
1

215 K sink
0.9
0 K sink

0.8

0.7
radiator flow fraction

0.6
Tsetpoint = 283 K
0.5 ΔTdesign 20 K
T∞ = 215 K
0.4

0.3

0.2

0.1

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
load fraction

Figure 2 – Radiator Bypass System Radiator Flow

5
Figure 2 also shows the change in radiator flow fraction with heat load in a 0 K
sink. At design heat load, the radiator flow is approximately 30% and it drops off
non-linearly as the heat load is decreased.
Figure 3 shows the radiator outlet temperature as a function of heat load fraction
for the two heat sinks. The freezing points of several heat transfer fluids are also
noted on the graph. The figure shows that an ammonia external loop requires
more than 60% of the design heat load to avoid radiator freezing in a deep space
environment. Even with Freon-212, which is a less desirable working fluid owing
to its poor heat transfer properties, the heat load must be maintained at more
than 25% of the design heat load to avoid freezing. The freezing point of
HFC-245fa is also noted in the graph. It is a “green” analog of the Freon-21 that
is used in the Space Shuttle Orbiter external ATCS and has a lower freezing
point than ammonia.

300

250
radiator outlet temperature (K)

200 NH3 195 K

HFC-245fa 171 K

Freon-21 138 K
150

215 K sink
0 K sink
100

Tsetpoint = 283 K
ΔTdesign 20 K
50
T∞ = 215 K

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
load fraction

Figure 3 – Radiator Bypass System Radiator Outlet Temperature

The minimum required heat load required in a radiator bypass system is much
larger than the heat rejection if the entire radiator were at the fluid freezing point
(the absolute minimum required heat load) because:
ƒ the cold environment causes a cold radiator return temperature.

2
Freon-21 is no longer manufactured owing to its high ozone depletion potential; however NASA
has stockpiles for use on the Space Shuttle Orbiter ATCS that could be made available for future
spacecraft.

6
 ƒ cold radiator return temperature causes the radiator bypass valve to open,
mixing warm radiator inlet flow with the return flow to maintain the required
mix temperature.
ƒ the resulting decreased radiator flow further decreases the radiator return
temperature which requires the bypass valve to open further.
This positive feedback loop results in very low flows through the radiator in cold
environments. In fact, at the freeze limit only a small fraction of the total loop
flow is passing through the radiator. These low flows require relatively high heat
loads to avoid freezing.

Regenerator Bypass Setpoint Control ATCS

Overview - To address the issue of high required minimum system heat load
during deep space transit, a novel setpoint temperature control methodology was
conceived. Rather than using a radiator bypass valve, a regenerative heat
exchanger is placed inboard of the radiators. A cold side bypass valve is used to
set the return temperature. The resulting loop architecture is shown in Figure 4.

7
 Pressurized Volume
coldplate

coldplate

Accum.
H/X

heater
pump

interloop
H/X

Accum.

coldplate coldplate
pump

T
regen
H/X

radiators

Figure 4 –Active Thermal Control Design with Regenerator Bypass

During operation, the regenerator bypass flow is varied, mixing cold radiator
return fluid and warm regenerator outlet fluid to maintain the system setpoint.
The regenerator is fully bypassed at the highest heat load in the design
environment, sending all the flow to the radiator - just as in the traditional two
loop system. At the lowest low heat load for stable operation, the bypass flow is
closed off – sending all the flow through the regenerator. This lowers the radiator
inlet temperature well below the system setpoint while maintaining full flow
through the radiators. By dramatically decreasing the average radiator
temperature at low heat loads, this concept substantially reduces the heat load
required to avoid freezing.
System Analysis - The radiator model described in the previous section was
modified to explore the performance of the Regenerator Bypass ATCS. The
model was modified to include the regenerator and energy balances were
included across key components.

8
The key temperatures are identical to the traditional two loop analysis:
ƒ The setpoint temperature is 283 K
o here it is the regenerator mix temperature
ƒ The external loop temperature drop is 20 K at the design point
o maximum load in design sink temperature – full regenerator bypass
ƒ The design sink temperature is 215 K
Because the same design conditions are used as for the two loop system, the
value of C is also the same.
Off-design cases were analyzed for the minimum required heat load by assuming
that the regenerator bypass was closed. The radiator flow fraction was unity and
the required radiator inlet temperature was calculated for a given outlet
temperature using eqn (6). This yielded the heat load fraction, Fload,min, from a
simple ratio
Tradiator,in − Tradiator,out
Fload,min = (9)
ΔTdesign
Figure 5 shows the minimum radiator temperature as a function of heat load. .
The figure shows that heat loads as low as 1/3rd are possible for an ammonia
system operating in a deep space sink. If Freon-21 is used as the working fluid,
heat loads less than 10% of full load are achievable.
200
NH3 195 K

190

180

HFC-245fa 171 K
T minimum (F)

170

160

Tsetpoint = 283 K
150 ΔTdesign 20 K
T∞ = 215 K

140

Freon-21 138 K
130
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
load fraction

Figure 5 – Highest Minimum Radiator Temperature at Various Heat Loads

Regenerator Bypass System – T sink = 0 K

9
The required regenerator performance can be calculated from its heat load and
internal temperature difference. The required design heat load, Qregen,design, is
& max c p (Tsetpo int − Tradiator,out )
Qregen,design = m (10)

The loop design heat load, Qloop,design, is defined as

Qloop,design = m
& max c p ΔTdesign (11)

so
(Tsetpo int − Tradiator,out )
Q regen,design = Q loop,design (12)
ΔTdesign

Since (Tsetpoint -Tradiator,out)/ΔTdesign is greater than unity, the regenerator heat

transfer will always be a multiple of the system design heat load.

The regenerator internal temperature difference at zero bypass is

Tradiator,in - Tradiator,out, so, using eqns. (9) and (10), the required regenerator
performance, UAregen,design, is
& c p (Tsetpo int − Tradiator ,out )
m
UA regen,design = (13)
Fload ΔTdesign
or, including eqn. (11)
UA regen,design (Tset − Tradiator ,out )
= (14)
Qloop,design Fload ΔTdesign
2

Figure 6 shows the regenerator performance required to achieve the desired

minimum radiator temperatures. The regenerator performance is given as
UAregenerator/Qloop,design. As the desired radiator outlet temperature decreases, the
minimum heat load fraction decreases (as shown in Figure 6). This decreases
the driving temperature difference across the heat exchanger. These two factors
work together geometrically to increase the required regenerator performance at
lower minumum heat loads.

10
 100000

Tsetpoint = 283 K
HFC-245fa 171 K
ΔTdesign 20 K
T∞ = 215 K
regenerator capability (W/K)

NH3 195 K

10000

Shuttle GSE H/X

0.18x0.08x0.33 m
Freon-21 both sides
UA=2600 W/K
Freon-21 138 K

1000
130 140 150 160 170 180 190 200
T minimum (K)

Figure 6 – Regenerator Sizing

Regenerator Bypass System – T sink = 0 K

Figure 6 shows that, for a 3 kW ammonia external system, a regenerator similar

in size to the Space Shuttle ground support heat exchanger (GSE H/X) is
sufficient – even before accounting for the superior heat transfer properties of
ammonia vs. Freon-21.

System Comparison
Table 1 shows a comparison of the two two-loop systems. The minimum
allowable temperatures are taken as the working fluid freezing point plus a 5 K
margin. The table shows that the regenerator bypass system yields a much
larger performance range for a given set of system design specifications. It
shows that, for each of the commonly used heat transfer fluids, the regenerator
bypass system can operate at less than half of the minimum heat load of the
traditional radiator bypass system. On a case by case basis, the additional mass
of the regenerator may be worthwhile for the design flexibility that it offers. The
table also shows that an ammonia regenerator system has a comparable
turndown to a traditional Freon-21 system. With similar requirements, the
superior performance of ammonia as a heat transfer fluid will likely outweigh the
additional mass of a regenerator.

11
 Regenerator Regenerator
Traditional Size
Bypass
Two-Loop UAregenerator
System ------------------------
min load Qdesign
min. load
fraction
fraction (1/K)
Ammonia
Tfreeze = 195 K 0.65 0.33 650
Tminimum = 200 K
HFC-245fa
Tfreeze = 171 K 0.48 0.20 1400
Tminimum = 176 K
Freon-21
Tfreeze = 138 K 0.29 0.08 5100
Tminimum = 143 K
Table 1 – Loop Performance Comparison

Conclusion
A new methodology for spacecraft active thermal control system setpoint control
has been described and analyzed. The regenerator bypass system offers a
greatly expanded allowable system heat load turndown for a given working fluid
by replacing the radiator bypass with a regenerator and bypass. It also allows
the substitution of higher performance working fluids for the same design
parameters. This ATCS architecture holds promise for decreasing system mass
and for increasing operational flexibility.

12